obsolete's SE/30 Reloaded Build

[obsolete]

So, build update. Here's the current state of the board:


Pretty close to done. Things left on the to-do list:

• Waiting for the PLCC to QFP carrier boards from OSH Park for the ADB PIC16
• Solder pins on the Bourns Again filters and install those
• Populate the RTC and surrounding components
• Install the PDS slot
• Oh yeah, NMI and reset switches, need those too; going to try 3D printing the little supports for them
• Waiting for Digi-Key to deliver 20-pin DIP sockets for the PALs

The reason for that last one is that I bought DIP sockets from Phoenix Enterprises and they are JUNK! The plating on them is just so bad, they are almost unsolderable. It's not that the legs are oxidized, but when soldered, the tin just disappears and exposes bare copper that I can't get my solder to wet. I'm just using Kester 44, nothing too weird, and it works fine on everything else.

I also bought the video ROM socket, CPU socket, PLCC sockets, SCSI, floppy, system ROM, and RAM sockets from Phoenix and they are all fine. The PGA CPU socket had a few pins that were mildly oxidized and took a little extra flux and heat to wet, but that's understandable since these are NOS parts. But the HWS2078 20-pin DIP sockets I would recommend staying far away from.

Speaking of sockets, this was my first time installing PLCC sockets. Based on what I'd seen people on YouTube do, either cutting out the bottom of the socket or using solder paste, I was a little worried about installing these, but it didn't end up being bad at all. A fine tipped iron, flux, and patience, and I found it pretty easy to get into a groove of soldering the pins by hand. I'm happy with how they came out.

 

[croissantking]

Why did you use PLCC sockets for just the GLUE, FPU and UH7?

 

[obsolete]

Why did you use PLCC sockets for just the GLUE, FPU and UH7?

GLUE: I wasn't 100% confident in my ability to remove it from the donor without damage, and if it ended up not working, that 84-pin PLCC was the one chip on the board that I didn't ever want to have to desolder from my new board. In hindsight, removing it from my donor board wasn't that bad because I ended up buying one of these nozzles and it worked great.
FPU: Second largest PLCC, second largest pain to desolder if it ended up being bad. I bought a cheap one from random eBay seller in China, so I want to be prepared in case it doesn't work.
UH7: I want all the PALs to be socketed so I can play with programming new ones at some point.

 

[zigzagjoe]

So, build update. Here's the current state of the board:
View attachment 69264

Pretty close to done. Things left on the to-do list:

• Waiting for the PLCC to QFP carrier boards from OSH Park for the ADB PIC16
• Solder pins on the Bourns Again filters and install those
• Populate the RTC and surrounding components
• Install the PDS slot
• Oh yeah, NMI and reset switches, need those too; going to try 3D printing the little supports for them
• Waiting for Digi-Key to deliver 20-pin DIP sockets for the PALs

The reason for that last one is that I bought DIP sockets from Phoenix Enterprises and they are JUNK! The plating on them is just so bad, they are almost unsolderable. It's not that the legs are oxidized, but when soldered, the tin just disappears and exposes bare copper that I can't get my solder to wet. I'm just using Kester 44, nothing too weird, and it works fine on everything else.

I also bought the video ROM socket, CPU socket, PLCC sockets, SCSI, floppy, system ROM, and RAM sockets from Phoenix and they are all fine. The PGA CPU socket had a few pins that were mildly oxidized and took a little extra flux and heat to wet, but that's understandable since these are NOS parts. But the HWS2078 20-pin DIP sockets I would recommend staying far away from.

Speaking of sockets, this was my first time installing PLCC sockets. Based on what I'd seen people on YouTube do, either cutting out the bottom of the socket or using solder paste, I was a little worried about installing these, but it didn't end up being bad at all. A fine tipped iron, flux, and patience, and I found it pretty easy to get into a groove of soldering the pins by hand. I'm happy with how they came out.
View attachment 69265

That's fantastic work hand soldering those with an iron - no melted plastic at all.

 

[robin-fo]

Wow! This looks incredibly nice!

 

[obsolete]

Thanks guys! Just wanted to show that it's possible to solder these cleanly without busting the center support out or using solder paste, which were the only two examples I saw before trying it myself.

I'm using a microscope, of course; I wish I had a trinocular microscope so I could make fancy videos but the best I can do with my current setup is stick my smartphone camera down one of the eyepieces and take photos.

The technique that worked well for me was to apply a thin film of liquid flux to the pads on the board with a cotton swab. Don't drown it, just a thin film. I'm using a small chisel tip iron. It's a Metcal, but pretty much anything will work for small stuff like this because there isn't a lot of thermal mass you're fighting against. The chisel shape is nice because it will hold a small bead of solder at a predictable spot on the tip. Any standard fine tip can be ground into this shape, but you'll have to re-tin it (and mind the lead dust!).

I know that most of us were taught "don't heat the solder, heat the work" but we can make a little exception in this case . I dab a little solder on the iron so that there's a small bead of molten solder sitting on the tip, then touch it where the socket leg and the PCB pad meet. Thanks to the liquid flux, the solder will flow and wet the whole pad and the socket leg and form a smooth, shiny joint. After tacking two corners, I just worked my way around the perimeter, dab, dab, dab, dab. I did all the ones that I could approach from one side first; because of the plastic supports, some legs can only be approached from the left or the right, and I got tired of having to switch positions back and forth, so once I had the socket tacked, I would do all the lefts, then all the rights, and I was able to get into a nice groove and make smooth, consistent joints pretty quickly this way.

 

[Kouzui]

So, build update. Here's the current state of the board:

Looking good! I wouldn't have bothered with the sockets, but that's just because they seem like too much trouble to me.

 

[obsolete]

The board is pretty much finished, and I'll be testing it this weekend! I'll post pictures and hopefully have exciting news to share later.

I got the QFN to PLCC adapters for the ADB chips this week. Thanks for the design, max1zzz!



So OSH Park says they don't guarantee castellated vias, and sure enough, only 3 out of the 4 sides of the adapters turned out okay. The side missing the via barrels is the one with the most pins to solder, and it's the same on all 3 parts! Oh well, just a little extra hassle, but I'll get it done. I have the ADB controller from the donor populated on the new board already, just to start with a known quantity before throwing more "aftermarket" parts on there, so it'll be a bit longer before I have this one programmed and installed.

Does anyone want the other two adapter boards? Like I said, all three are identical, down to the missing via barrels on the side with 7 pads to solder. I could program and solder PICs onto them too.

 

[croissantking]

Does anyone want the other two adapter boards? Like I said, all three are identical, down to the missing via barrels on the side with 7 pads to solder. I could program and solder PICs onto them too.

If I were more local to you I'd grab 'em, I really like the idea of a Reloaded board that doesn't use original parts.

Interested to see how they would attach to the SE/30 Reloaded PCB, I can't quite see how it would work at the moment with those PLCC pads.

 

[obsolete]

Well, I've run into trouble. For some reason, I am not seeing C16M coming out of the GLUE chip. No 16MHz clock means pretty much nothing works; the CPU is not running.

I posted a separate troubleshooting thread here: https://68kmla.org/bb/index.php?threads/troubleshooting-glue-c32m-and-c16m.46886/

 

[croissantking]

Seems like ATF16V8Bs should work, no?

To confirm, they do. I’ve tested one out.

Link to part

 

[obsolete]

To confirm, they do. I’ve tested one out.

Link to part

Yep, I used ATFs for all my PALs and have confirmed that they all work on an original Apple logic board. I remember reading elsewhere about some issues programming them, or needing to program them multiple times for them to work. I read the datasheet and the max programming voltage is 14V. The default programming voltage in the XGPro software for my T48 programmer was 12V. I decided to program them at 13V, and they all worked well the first time.